OpenAI Claims Internal AI Model Disproved Exponential Time Hypothesis

27 Oct 2026, 9:58 am · 2h ago · 1 min read · OfficeChaiNew

New research from OpenAI claims that one of its internal AI models has disproved the Exponential Time Hypothesis (ETH), a key cornerstone of computational complexity theory. Formulated in 2001, the hypothesis asserts that core NP-complete problems like 3-SAT cannot be solved in sub-exponential time. An algorithmic breakthrough disproving ETH would have sweeping consequences for algorithm optimization, computational theory, and standard cryptographic protocols. While OpenAI has shared preliminary findings from its internal model, the wider academic and computer science community awaits complete technical papers and formal mathematical proofs to independently validate the breakthrough.